Optimizing the social force model using new hybrid WOABAT-IFDO in crowd evacuation in panic situation.

Abstract

This paper addresses the need for improvement in the Social Force Model (SFM) crowd evacuation model in the context of egress studies and current emergency research. As the current classical evacuation model, the Social Force Model lacks decision-making ability for finding the best directions towards an exit. Crowd searching for route choices in crowd evacuation simulations for panic situations remains inaccurate and unrealistic. There is a need for SFM to be incorporated with an intelligent approach in a simulation environment by adding in behaviour of following the position indicator to guide agents towards the exit to ensure minimal evacuation time. Congestion in pedestrian crowds is a critical issue for evacuation management, due to a lack of or lower presence of obstacles. Thus, this research proposes optimization using the one of the latest nature inspired algorithm namely WOABAT-IFDO (Whale-Bat and Improved Fitness-Dependent Optimization) in the SFM interaction component. Optimization takes place by randomly allocating the best position of guide indicator as an aid to the for better evacuation time and exploring the dynamics of obstacle-non obstacle scenarios that can disperse clogging behavior with different set of agent’s number for better evacuation time and comparing it with single SFM simulation. Finally, validation is conducted based on the proposed crowd evacuation simulation time, which is further based on standard evacuation guidelines and statistical analysis methods.